Annually, over 300,000 US patients survive an acute critical illness but fail to recover sufficiently to permit ventilator liberation, a syndrome knownas prolonged mechanical ventilation (PMV). PMV is associated with high mortality, prolonged disability, and extraordinary resource utilization. Because these patients are incapacitated due to illness, their surrogates must make life support decisions on their behalf. However, the quality of the decision making process is seriously deficient and may lead to decisions that are inconsistent with a patient's preferences, prolonged life support that is extraordinarily costly an ineffective, and psychological distress among patients' surrogate decision makers. To address these problems, we developed a decision aid for PMV surrogate decision makers that promotes shared decision making between surrogate decision makers and clinicians. In a recent pilot study, we found that compared to usual care control, the decision aid improved surrogate-clinician concordance for prognosis, the quality of communication, and the likelihood that clinicians and surrogates engaged in life support discussions. It also improved surrogates' decisional conflict and trust in clinicians. To confirm and extend our promising results, we now propose a multicenter randomized controlled trial (RCT) to test the efficacy of the PMV decision aid vs. usual care control in improving: (1) decision making quality, defined as clinician-surrogate concordance for prognosis, quality of communication, and medical comprehension; (2) surrogates' psychological distress; and (3) patients' health care utilization. We aim to enroll 600 surrogate decision makers for PMV patients, as well as patients' ICU physicians and nurses, from five diverse centers that make up an existing PMV research network. We will use generalized linear models to test for differences in outcomes between decision aid and usual care control groups over a 6-month follow up period. The proposed study addresses key research needs in decision making highlighted by the NIH and the Institute of Medicine, and also meets a call for innovation in decision aids specified by the 2010 Affordable Care Act. Innovative aspects of this pragmatic intervention include: (a) its use of web-based tablet computer technology to address surrogates' informational needs; (b) its function as a family meeting adjunct to promote individualized shared decision making, (c) the targeting of communication-associated risk factors for psychological distress, (d) its potential for inexpensive dissemination and adaptation to broader populations, (f) its ability to accommodate updates in component clinical risk estimates, and (g) the inclusion of an economic evaluation of effect. Overall, this research could have a significant, sustained impact on critical care's future clinica and research approach.